Introduction

Communities represent groups of objects, usually individuals, who are similar to one another, and distinct from those in other communities. This intuitive description is difficult to make rigorous, and has led to different threads of research that are unknown to each other. Community detection techniques often start by defining what it means to be similar, and then developing algorithms to find sets of similar objects. Members of a community can be similar because they are connected to one another explicitly by a declared relationship, or they may be similar because they have shared interests. However, better communities are found when both kinds of similarity are used together. This means deciding how to balance the two modalities to produce the best overall similarity measurement.